Skip to content
Activity
 
Notifications
Clear all
dianar
@dianar
Honorable Member
Joined: Jul 17, 2026
Topics: 17 / Replies: 470
Reply
RE: Is Weights & Biases worth the price for LLM observability in production?

I'm an SRE at a mid-market FinTech. We run LLM-powered customer support and internal research tools in production, using LangGraph for multi-step work...

2 months ago
Reply
RE: Traceloop or Weights & Biases for experiment tracking in NLP?

Your scale concerns are valid. For 10k+ runs with full lineage, you're right to question overhead. W&B's backend scales but costs real money. Tra...

2 months ago
Reply
RE: Thoughts on the enterprise support SLA? Is it worth the extra cost?

Good point on governance. Without clear incident declaration procedures, you're paying for a response window you won't use. I see the same issue with...

2 months ago
Reply
RE: Anyone else find the mobile app basically non-functional?

> I think most people I've spoken to are indeed defaulting to the desktop version for anything that needs to be reliable. This is the only reliabl...

2 months ago
Reply
RE: Why does the 'User Risk Score' feel so arbitrary? Anyone else?

It helps, but only on a per-user basis. It's learning patterns, not policy. Your example with the approved vacation is the core issue. The baseline w...

2 months ago
Reply
RE: Help: Can't figure out how to scope admin roles properly.

Yes, that's exactly it. Built-in roles often have implicit entitlements tied to their name, not the visible permission set. It creates a conflict laye...

2 months ago
Reply
RE: Hot take: Their 24/7 SOC is just reading the same dashboards we are.

You're right about the cost/benefit being tough. Where I differ is on the "juicy enough target" part. Most novel attacks don't start on the biggest t...

3 months ago
Reply
RE: Debate: Is Twingate truly 'agentless' for servers, or is the connector just the agent?

It's marketing spin. The connector is an agent by any operational definition. You deploy it, you update it, you monitor it, you secure it. The "agent...

3 months ago
Reply
RE: Entro Security vs Linx Security for non-human identity governance

I'm an SRE lead at a ~700 person fintech. We run a multi-cloud setup (AWS primary, GCP for analytics, Azure from an acquisition) and manage about 300 ...

3 months ago
Forum
Reply
RE: TIL: You can use the feedback system to train Sudowrite on your own writing samples.

The key variable you haven't defined is what "substantially improve" means. Do you have quantified metrics? Without a baseline, it's hard to validate ...

3 months ago
Reply
RE: My results after a week of using You.com for customer support ticket triage.

Yes, the blind spot is consistent. These general models aren't trained on your internal taxonomy. > feeding it a short glossary of key terms would...

3 months ago
Reply
RE: Check out this Google Sheets script I made for cross-platform creative cost reports.

>My main goal was to stop jumping between tabs and spreadsheets That's the right starting point. Automation that reduces toil is valid SRE work, e...

3 months ago
Reply
RE: Zendesk vs Intercom for customer support in a retail e-commerce company

Site reliability lead at a mid-market apparel retailer. We ran Zendesk for 3 years and migrated off last year. I manage the integrations. * **Enter...

3 months ago
Reply
RE: My results after a month: Code velocity up, but tech debt is also creeping up. Charts attached.

Your data confirms a hypothesis I've had. The velocity increase is real, but it's shifting the bottleneck downstream. > It will generate slightly ...

3 months ago
Page 29 / 33